.png)
AI Asset Search with Zeet fully-managed Vector Databases
.svg)
.svg)
Abstract
Mirage lets users perform an image or video search like they would on Google, but it goes a step further and actually can take complex queries, which it learns from as you search. Instead of typing in something and sifting through results, you can type in exactly what you’re looking for, and Mirage will effectively sift through the results for you. If you don’t see what you’re looking for, you can use Mirage to generate an image, or provide a similar image which Mirage will use to narrow your search.
Mirage relies heavily on their Vector Database for every part of their product. From smart-queries, to image gen and reference matching, to learning your preferences, their vector database is a core component of their stack.
Mirage used Zeet Blueprints to deploy Vector Database via Helm, Redis, NodeJS Express Application via Git repo, and an Inference Service deployed via Docker.
"We had built our internal infrastructure in house, through a combination of multiple cloud providers (AWS & GCP) and third party services. As we expanded to B2B and enterprise clients it became critical for us to create tenancy for each customer while also providing the option to deploy into a customer's cloud. Zeet helps us instantly clone our infrastructure and easily manage deployments & updates regardless of the cloud provider. Now, we don't have have to worry about the complexities of managing that ourselves and can focus on our application code."
